Case study

How a support team resolved 15% more enquiries per hour with AI support

Software company · customer support · 5,172 people

The problem

Practised employees are considerably faster than new or less experienced ones.

In a team this size, onboarding decides output: whoever knows the cases settles them in minutes, whoever is new goes searching. And training costs time from exactly the people you can least afford to pull off the floor.

What was built

An assistant that suggests the next sentence on every case.

Employees are helped directly inside the case. The assistant reads along and suggests answers based on what experienced colleagues wrote in comparable cases. It was the employee who sent the answer.

The result

15% more enquiries resolved per hour.

  • 15%more enquiries resolved per hour
  • 5,172people in the study
  • New employeesshow the biggest gain in output

The effect was strongest for those who had just started: they got faster, and their answers got better. The most practised gained pace as well. The tool makes output more even across everyone.

Human in the loop: here the AI only suggests, while the employee decides what gets sent.

What fits

How DIGITLZ can help you

AI training for your team

Your team learns to work with AI inside the workflow and to judge results instead of believing them. Up to 10 people per session.

To the training

Corporate Context

The assistant was only good because it knew the company's own cases. That takes a context your team and your agents can read.

To Corporate Context

Peer-reviewed study, published in the Quarterly Journal of Economics. The assistant was rolled out in stages, and the comparison ran against the colleagues who did not have it yet.

Source: Brynjolfsson, Li & Raymond, Stanford / MIT, Quarterly Journal of Economics 140(2), 2025, pp. 889–942.

Your case

Where would your team gain the most?

Book an intro call